Repository files navigation

SIFT	DISK	IMAGES ORIENTATION	DENSE WITH ROMA

SIFT	SUPERGLUE

Multivew matcher for SfM software. Support both deep-learning based and hand-crafted local features and matchers and export keypoints and matches directly in a COLMAP database or to Agisoft Metashape by importing the reconstruction in Bundler format. Now, it supports both OpenMVG and MicMac. Feel free to collaborate!

Whiledev branch is more frequently updated,master is the default more stable branch and is updated fromdev less frequently. If you are looking for the newest developments, please switch todev.

For how to use DIM, check theDocumentation (updated for the master branch).

Please, note thatdeep-image-matching is under active development and it is still in an experimental stage. If you find any bug, please open an issue.For the licence of individual local features and matchers please refer to the authors' original projects.

Key features:

• Multiview
• Large format images
• SOTA deep-learning and hand-crafted features
• Support for image rotations
• Compatibility with several SfM software
• Support image retrieval with deep-learning local features

Want to run on a sample dataset? ➡️

Want to run on your images? ➡️

DIM can also be utilized as a library instead of being executed through the Command Line Interface (refer to theUsage Instructions).

For quick examples, see:

• demo.py - Simple script demonstrating the basic workflow
• demo.ipynb - Interactive notebook version of the demo
• notebooks/sfm_pipeline.ipynb - Complete SfM pipeline with detailed explanations

For installing deep-image-matching, we recommend usinguv for fast and reliable package management:

# Install uv if you haven't alreadycurl -LsSf https://astral.sh/uv/install.sh| sh# Create and activate a virtual environmentuv venv --python 3.9source .venv/bin/activate# On Windows: .venv\Scripts\activate

Then, you can install deep-image-matching using uv:

uv pip install -e.

This command will install the package in editable mode, allowing you to modify the source code and see changes immediately without needing to reinstall. If you want to use deep-image-matching as a non-editable library, you can also install it without the-e flag.

This will also installpycolmap as a dependency, which is required for running the 3D reconstruction.If you have any issues withpycolmap, you can manually install it following the official instructionshere.

To verify that deep-image-matching is correctly installed, you can try to import the package in a Python shell:

importdeep_image_matchingasdim

To test most of the functionality, run the tests to check if deep-image-matching is correctly installed, run:

uv run pytests

For more information, check thedocumentation.

This project has migrated from conda/pip touv for dependency management. Benefits include:

• Faster installation: uv is significantly faster than pip for dependency resolution and installation
• Better dependency resolution: More reliable resolution of complex dependency trees
• Lockfile support:uv.lock ensures reproducible installations across different environments
• Integrated tooling: Built-in support for virtual environments, Python version management, and project building
• Cross-platform consistency: Better support for different operating systems and architectures

If you have any issue with uv, you prefer to have a global installation of DIM, or you have any other problem with the installation, you can use conda/manba to create an environment and install DIM from source using pip:

git clone https://github.com/3DOM-FBK/deep-image-matching.gitcd deep-image-matchingconda create -n deep-image-matching python=3.9conda activate deep-image-matchingpip install -e.

For Docker installation, see theDocker Installation section in the documentation.

For a quick start, check out thedemo.py script ordemo.ipynb notebook that demonstrate basic usage with the example dataset:

python demo.py --dir assets/example_cyprus --pipeline superpoint+lightglue

The demo runs the complete pipeline from feature extraction to 3D reconstruction using the provided example dataset.

A similar demo example is also available as a notebook indemo.ipynb.

Use the following command to see all the available options from the CLI:

python -m deep_image_matching --help

For example, to run the matching with SuperPoint and LightGlue on the example_cyprus dataset:

python -m deep_image_matching --dir assets/example_cyprus --pipeline superpoint+lightglue

The--dir parameter defines the processing directory, where all the results will be saved. This directory must contain a subfolder namedimages with all the images to be processed.

Deep-image-matching can also be used as a Python library. For a comprehensive example showing the complete SfM pipeline, seenotebooks/sfm_pipeline.ipynb.

For detailed usage instructions and configurations, refer to thedocumentation.

For advanced usage, please refer to thedocumentation and/or check thescripts directory.

To run the matching with different local features and/or matchers and marging together the results, you can use scripts in the./scripts directory for merging the COLMAP databases.

python ./join_databases.py --helppython ./join_databases.py --input path/to/dir/with/databases --output path/to/output/dir

To export the solution to Metashape, you can export the COLMAP database to Bundler format and then import it into Metashape.This can be done from Metashape GUI, by first importing the images and then use the functionImport Cameras (File -> Import -> Import Cameras) to select Bundler file (e.g., bundler.out) and the image list file (e.g., bundler_list.txt).

Alternatevely, you can use theexport_to_metashape.py script to automatically create a Metashape project from a reconstruction saved in Bundler format.The scriptexport_to_metashape.py takes as input the solution in Bundler format and the images and it exports the solution to Metashape.It requires to install Metashape as a Python module in your environment and to have a valid license.Please, refer to the instructions athttps://github.com/franioli/metashape.

Any contribution to this repo is really welcome!If you want to contribute to the project, please, check thecontributing guidelines.

See theTODO list for the list of features and improvements that are planned for the future.

If you find the repository useful for your work consider citing the papers:

@article{morelli2024_deep_image_matching,AUTHOR ={Morelli, L. and Ioli, F. and Maiwald, F. and Mazzacca, G. and Menna, F. and Remondino, F.},TITLE ={DEEP-IMAGE-MATCHING: A TOOLBOX FOR MULTIVIEW IMAGE MATCHING OF COMPLEX SCENARIOS},JOURNAL ={The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences},VOLUME ={XLVIII-2/W4-2024},YEAR ={2024},PAGES ={309--316},DOI ={10.5194/isprs-archives-XLVIII-2-W4-2024-309-2024}}

@article{morelli2022photogrammetry,title={PHOTOGRAMMETRY NOW AND THEN--FROM HAND-CRAFTED TO DEEP-LEARNING TIE POINTS--},author={Morelli, Luca and Bellavia, Fabio and Menna, Fabio and Remondino, Fabio},journal={The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences},volume={48},pages={163--170},year={2022},publisher={Copernicus GmbH}}

@article{ioli2024,title={Deep Learning Low-cost Photogrammetry for 4D Short-term GlacierDynamics Monitoring},author={Ioli, Francesco and Dematteis, Nicolò and Giordan, Daniele and Nex, Francesco and Pinto Livio},journal={PFG – Journal of Photogrammetry, Remote Sensing and Geoinformation Science},year={2024},DOI ={10.1007/s41064-023-00272-w}}

Depending on the options used, consider citing the corresponding work of:

• KORNIA
• HLOC
• COLMAP
• OpenMVG
• MICMAC
• used local features and matchers